This invention relates generally to identifying targeting criteria for content, and more particularly to automatically and dynamically identifying refined targeting criteria based on user responses to the content.
It is advantageous for content providers in online systems, such as social networking systems, to identify targeting criteria that will lead to increased interaction by users with the content or increased conversion resulting from content. Content providers commonly provide targeting criteria along with the content itself to online systems for presentation of the content to users. The initial targeting criteria provided by the content provider typically include information, such as age, gender, and region of residence for a user, or may include specific targeting time ranges for the content. Often times, the targeting criteria specified by the content provider may be fairly basic, static information, and may fail to include other more dynamic targeting criteria that result in effective responses for the content.